einr wrote:

The drives don't spin up,

If you don't connect anything else to the PSU, but a known working hard disk (I mean really nothing, not even the motherboard, only one molex -> HDD) does the drive spin?
If not, focus on the PSU before tossing the motherboard.

konc wrote:

einr wrote:

The drives don't spin up,

If you don't connect anything else to the PSU, but a known working hard disk (I mean really nothing, not even the motherboard, only one molex -> HDD) does the drive spin?
If not, focus on the PSU before tossing the motherboard.

Yeah, I've tried that, and the drive does spin up if it is connected to the PSU only. If I plug it into the controller card, there's no sign of life from anything. Only the CPU and PSU fans spin up, nothing else.

einr wrote:

am I correct in assuming this is the tag RAM chip pictured below?

Yes, that's the TAG RAM. Dunno if it's the correct size for your new board, though.

Weird that the hard drive doesn't spin up when connected to the controller 😕

oerk wrote:

einr wrote:

am I correct in assuming this is the tag RAM chip pictured below?

Yes, that's the TAG RAM. Dunno if it's the correct size for your new board, though.

Thanks for the confirmation!

Does the correct size/type of the tag RAM chip depend on anything other than the size of the installed cache? I'm thinking I'll replace all the cache chips from the new (MTI) board with the ones from the dead board.

What about 32-pin vs 28-pin cache chips? Are they interchangeable? Looks like the dead board has 8 x 28-pin chips and the new one has 4 x 32-pin...

28-pin SRAM chips top at 32kB each (not enough address pins in a 28-pin package to use more) that's why you need 8 of them to get 256kB. They are pin-compatible with the 32-pin sockets but you will have only 4 of those so if you reuse the 28-pin chips the max cache you can get is 128kB.

32-pin SRAM chips due to more address lines available on the package can go up to 512kB per chip.

einr wrote:

If I plug it into the controller card, there's no sign of life from anything.

Are you sure the controller card works?

Sometimes I get those symptoms on a working board if I use RAM that it doesn't like, or place the SIMMs in slots in a way that it doesn't like, or configure the RAM related jumpers in a way that it doesn't like, or connect a graphics card that it doesn't like. Some of these 486 boards are very finicky and you can't always count on the absence of beeps and other noises as guarantee of a dead board, although in your case that seems the most likely scenario.

devius wrote:

einr wrote:

If I plug it into the controller card, there's no sign of life from anything.

Are you sure the controller card works?

Sometimes I get those symptoms on a working board if I use RAM that it doesn't like, or place the SIMMs in slots in a way that it doesn't like, or configure the RAM related jumpers in a way that it doesn't like, or connect a graphics card that it doesn't like. Some of these 486 boards are very finicky and you can't always count on the absence of beeps and other noises as guarantee of a dead board, although in your case that seems the most likely scenario.

Not sure the controller card, or graphics card, or almost anything else works. I have no other suitable ISA or VLB cards to try, I'm afraid. The thing is, all the configuration (jumpers, RAM, etc) *should* be good since this was a fully assembled system when I got it. But it had also spent ten years on a damp garage floor, so anything/everything could have gone bad since it was last used.

That's always the most likely scenario. If everything else fails and all hope is lost you could try baking it in the oven. Some people had success with that technique, but personally I never tried it.

^^^ I've actually already tried baking the QDI board in the oven, no luck unfortunately.

Anyway, the new board arrived and the system works totally fine with it! The board came with a 486 DX/33 and 8 Mb of RAM installed so that's what I'm using at the moment since it's a known good configuration. Will update later to the DX4 that came with the first board.

The VGA card works, the multi I/O controller works, the hard disk works. It boots into DOS 6.22 and Windows 3.11 for Workgroups. Seems to have been someone's home computer, lots of Word and Works documents and even old scanned pictures of normal family stuff. System was last used in 1999. It's always kind of a weird feeling to dig through someone's old stuff.

Anyway, the cache issue. The system reports 256K L2 cache on boot, but that's what I expected since apparently all MTI boards do so, whether the cache is real or not. I transferred CACHECHK to the system and ran it. Got this output:

1 CACHECHK v4 2/7/96  Copyright (c) 1995 by Ray Van Tassle. (-h for help)
2 CMOS reports: conv_mem= 640K, ext_mem= 7,168K, Total RAM= 7,808K
3 Clocked at 486 33.4 MHz
4 Reading from memory.  
5 MegaByte#:      --------- Memory Access Block sizes (KB)-----
6       1    2    4    8   16   32   64  128  256  512 1024 2048 4096 <-- KB
7 0:   31   31   31   32   45   45   45   45   45  127   --   --   --    æs/KB
8 4:   31   31   31   32   45   45   45   45   45  127  127  127  127    æs/KB
9 5  6  7  <--- same as above.
10
11 Extra tests----
12 Wrt  41   41   41   41   42   41   41   41   41   41   41   41   41<-Write mem
13 This machine seems to have both L1 and L2 cache. [read]
14    L1 cache is   8KB --  34.5 MB/s  30.4 ns/byte  (401%) (142%)   3.9 clks
15    L2 cache is 256KB --  24.2 MB/s  43.3 ns/byte  (281%) (100%)   5.5 clks
16    Main memory speed --   8.6 MB/s  121.3 ns/byte  (100%) [read]  15.5 clks
17    Effective RAM access time (read ) is 485ns (a RAM bank is 4 bytes wide). 
18    Effective RAM access time (write) is 157ns (a RAM bank is 4 bytes wide). 
19    Clocked at 486 33.4 MHz.  Cache ENABLED.
20 Options: -t0

So... That's good, right? Isn't CACHECHK pretty much impossible to fool? Do any of the numbers look weird to you guys, or is this a real bona-fide 256K cache system? Sure looks like it! 😎

Looks like a pretty solid indication of 256k.

This is one of many reasons to have a lot of spares on hand for testing but collecting these days is getting harder, it is possible that board might still be working and that it is being picky but it is odd that the rest of the system isn't working.